Visual signal detection. IV. Observer inconsistency.

Historically, human signal-detection responses have been assumed to be governed by external determinants (nature of the signal, the noise, and the task) and internal determinants. Variability in the internal determinants is commonly attributed to internal noise (often vaguely defined). We present a variety of experimental results that demonstrate observer inconsistency in performing noise-limited visual detection and discrimination tasks with repeated presentation of images. Our results can be interpreted by using a model that includes an internal-noise component that is directly proportional to image noise. This so-called induced internal-noise component dominates when external noise is easily visible. We demonstrate that decision-variable fluctuations lead to this type of internal noise. Given this induced internal-noise proportionality (sigma i/sigma 0 = 0.75 +/- 0.1), the upper limit to human visual signal-detection efficiency is 64% +/- 6%. This limit is consistent with a variety of results presented in earlier papers in this series.

Vertebral trabecular bone: comparison of single and dual-energy CT measurements with chemical analysis.

We have done detailed analysis of 50 excised L1 vertebral bones including single-energy CT, preprocessing dual-energy CT, as well as a number of physical and chemical analyses to determine the amount of fat and the elemental composition. We found a good correlation (0.91) between single-energy CT measurements and mineral element content and excellent correlation (0.95) between dual-energy CT measurements and mineral element content. The SEEs of the CT mineral were 14 and 10 mg (K2HPO4 equivalent) per cm3 for single and dual energy, respectively. We also report on the variation of fat and mineral content with respect to age, the quantitative effect of fat variation on CT values, and the relationship between calcium and phosphorus content.

Learned tolerance to the anticonvulsant effects of alcohol in rats.

Ethanol (1.5 g/kg, IP) administered to kindled rats blocked the seizures normally elicited in these subjects by electrical stimulation of the amygdala. Tolerance to this anticonvulsant effect developed following a series of ethanol injections only when the amygdaloid stimulation was administered during the periods of ethanol intoxication. Control subjects stimulated each day prior to ethanol administration displayed no tolerance to ethanol's anticonvulsant effects. Such findings emphasize the important role of learning in the development of alcohol tolerance and support the view that tolerance develops more rapidly for responses that are affected by the alcohol exposure.